Classifications

• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
  • H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
  • H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
• • G—PHYSICS
  • G10—MUSICAL INSTRUMENTS; ACOUSTICS
  • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
  • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
  • G10L19/002—Dynamic bit allocation
• • G—PHYSICS
  • G10—MUSICAL INSTRUMENTS; ACOUSTICS
  • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
  • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
• • G—PHYSICS
  • G10—MUSICAL INSTRUMENTS; ACOUSTICS
  • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
  • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
  • G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques

Definitions

• the present invention relates to an apparatus and method for encoding and decoding information signals, such as may occur in audio and video encoding, for example.
• the first method relies on predicting the time signal.
• the resulting Predictor error is then entropy-coded and sent for storage or transmission, eg in SHORTEN (see Tony Robinson: SHORTEN: Simple lossless and near lossless waveform compression Technical Report CUED / F-INFENG / TR.156, Cambridge University Engineering Department, December 1994) or AudioPaK (see Mat Hans, Ronald W. Schafer: Lossless Compression of Digital Audio, IEEE Signal Processing Magazine, June 2001).
• the second method uses as a first processing step a time-frequency transformation with subsequent lossy coding of the resulting spectrum.
• the error arising in the inverse transformation can then be entropy-coded in order to ensure lossless coding of the signal, e.g. LTAC
• MPEG-4 SLS Scalable Lossless Coding, see Ralf Geiger, et al .: ISO / IEC MPEG-4 High-Definition Scalable Advanced Audio Coding, 120th AES Convention, May 2006).
• the first possibility corresponds to a redundancy reduction.
• an uneven probability distribution of an underlying alphabet of the signal is exploited. For example, symbols that have a higher probability of occurrence are displayed with fewer bits than symbols with a smaller probability of occurrence.
• this principle is also called entropy coding.
• entropy coding During the encoding / decoding process, no data is lost. A perfect (lossless) reconstruction of the data is thus possible again.
• the second possibility concerns an irrelevance reduction. With this type of data reduction, information that is not relevant to the user is specifically removed. As a basis for this often models of natural perceptual limitations of the human senses are used.
• the input data are transformed from the time to the frequency domain and quantized there using a psychoacoustic model.
• this quantization introduces only enough quantization noise into the signal that it is imperceptible to the listener, but this can not be achieved for low bit rates - clearly audible coding artifacts are produced.
• the object of the present invention is thus to provide an alternative concept for coding and decoding information signals.
• the present invention is based on the finding that an information signal can be coded more cost-effectively if a sorting is carried out beforehand.
• an information signal or also an audio signal comprises a sequence of sampled values, wherein the sampled values can originate from a time or frequency signal, ie it can also be a sampled spectrum.
• sample is thus not meant to be limiting.
• a fundamental processing step may be to perform the sorting of an input signal according to its amplitude, which may also take place after any pre-processing has been performed.
• a time / frequency transformation, a prediction or a multi-channel redundancy reduction, for example in the case of multichannel signals may be carried out, in general also decorrelation methods.
• a possibly variable division of the signal into defined time segments, so-called frames can take place.
• a further division of these periods into subframes, which are then sorted individually, is possible.
• the sorted data and, on the other hand, a sorting rule which exists as a permutation of the indices of the original input values exist. Both data sets are then encoded as effectively as possible.
• Embodiments offer several possibilities, such as a prediction with subsequent entropy coding of the residual signal, i. determining prediction coefficients for a prediction filter and determining the residual signal as a difference between an output of the prediction filter and the input signal.
• curve fitting is carried out with suitable functional specifications and functional coefficients with subsequent entropy coding of the residual signal.
• a lossy coding done and thus eliminating the coding of the residual signal.
• Embodiments may also perform permutation coding, for example by constructing inversion tables and subsequent entropy coding, details of inversion tables being found, for example, in Donald E. Knuth: The Art of Computer Programming, Volume 3: Sorting and Searching, Addison-Wesley, 1998. NEN.
• a prediction of inversion tables and subsequent Entropiekodtechnik the residual signal can be carried out, as well as a prediction of the permutation and subsequent Entropiekodtechnik the residual signal.
• Embodiments may also achieve lossy coding by omitting the residual signal.
• combinatorial selection methods with subsequent numbering can be used in exemplary embodiments.
• Fig. Ia an embodiment of a device for coding
• Fig. Ib an embodiment of a device for decoding
• FIG. 2a shows an embodiment of a device for coding
• FIG. 2b shows an embodiment of a device for decoding
• 3a shows an embodiment of a device for coding
• 3b shows an embodiment of a device for decoding
• 4a shows an embodiment of a device for coding
• 4b shows an embodiment of a device for
• Fig. 6 shows an embodiment of an encoder
• Fig. 7 shows an embodiment of a decoder
• Fig. 12 shows an example of a sorted difference-coded signal and its residual signal
• FIG. 13 example of a sorted time signal
• FIG. 14 shows an example of sorted time values and a corresponding curve fitting
• Fig. 16 shows exemplary processing steps of most lossless audio compression algorithms
• 17 shows an embodiment of a structure of a prediction coding
• FIG. 18 shows an embodiment of a structure of a reconstruction in a prediction coding
• FIG. 19 shows an exemplary embodiment of warmup values of a prediction filter
• FIG. 20 shows an embodiment of a predictive model
• Fig. 21 is a block diagram of a structure of an LTAC encoder
• Fig. 22 is a block diagram of an MPEG-4 SLS encoder
• FIG. 23 shows a stereo-redundancy reduction after decorrelation of individual channels
• FIG. 24 shows a stereo-redundancy reduction before decorrelation of individual channels
• Fig. 25 is an illustration of the relationship between predictor ordering and total bit consumption
• FIG. 26 is an illustration of the relationship between quantization parameter g and total bit consumption
• FIG. Fig. 27 is an illustration of magnitude frequency response of a fixed predictor depending on its order p;
• 29 a to h show a representation of inversion tables in the 10th block (frame) of a noise-like piece
• FIGS. 30 a to h show a representation of inversion tables in the 20th block (frame) of a tonal piece
• 31 a and b show a representation of a permutation of a noise-like piece in the 10th block and a tonal piece, which has arisen from a sorting of time values;
• Fig. 32a shows a part of an audio signal, the corresponding permutation and inversion table LS, and 32b enlarges the permutation and the inversion table LS from the left picture;
• Fig. 33a enlarges a part of an audio signal, the corresponding permutation and inversion table LS, and Fig. 33b enlarges the permutation and the inversion table LS from the left picture;
• FIG. 34 shows a probability distribution (top) and a length of the codewords (bottom) of a prediction signal (fixed predictor) of an inversion table LB;
• FIG. 35a shows a probability distribution and Fig. 35b shows a length of codewords of a residual signal (below) of sorted time values produced by differential coding;
• FIG. 36 shows a percentage of a partial block decomposition with a least amount of data of a forward adaptive Rice coding over a residual signal of a fixed predictor of a piece including side information for parameters; the total block length is 1024 time values;
• FIG. 37 shows a percentage of a partial block decomposition with the lowest data volume of a forward adaptive Golomb coding over the residual signal of a fixed predictor of a piece including side information for parameters; the total block length is 1024 time values;
• Fig. 38 is an illustration of the operation of a history buffer
• Figures 39a and b show an illustration of the operation of an adaptation compared to an optimal parameter for the whole block
• FIG. 40 shows an embodiment of a forward adaptive arithmetic coding with the aid of a backward adaptive Rice coding
• FIG. 41 shows a representation of the influence of the block size on the compression factor F
• FIG. 42 shows a representation for lossless MS encoding
• FIG. 43 shows a further illustration for lossless MS encoding
• Fig. 44 is an illustration for selecting a best variation for stereo-redundancy reduction.
• identical or functionally identical functional elements have the same reference numerals and thus the descriptions of these functional elements in the various embodiments shown below are interchangeable.
• representative of discrete values of a signal is generally referred to as samples.
• sample is not intended to be limiting, samples may be obtained by sampling a time signal, a spectrum, a general information signal, and so forth.
• Fig. Ia shows an apparatus 100 for encoding a sequence of samples of an audio signal, each sample having a jump position within the sequence.
• the apparatus 100 includes means 110 for sorting the samples by size (after any preprocessing, eg, time / frequency transform, prediction, etc.) to obtain a sorted sequence of samples, each sample being one Has sorting position within the sorted sequence. Further, the apparatus 100 includes means 120 for encoding the sorted samples and information about a relationship between the source and sort positions of the samples.
• the apparatus 100 may further comprise preprocessing means configured to perform filtering, time / frequency transformation, prediction or multi-channel redundancy reduction to generate the sequence of samples.
• the means 120 for encoding may be configured to encode the information about the relationship between the source and sort positions as an index permutation.
• means 120 for encoding may include the information encode the relationship between the source and sort items as an inversion chart.
• the means 120 for coding may further be configured to code the sorted samples or the information about the relationship between the original and sorting positions with a difference and subsequent entropy coding or only an entropy coding.
• the device 120 may determine and encode coefficients of a prediction filter based on the sorted samples, a permutation, or an inversion table. In this case, furthermore, a residual signal which corresponds to a difference between the sampled values and an output signal of the prediction filter can be coded and enable lossless coding. The residual signal can be coded with an entropy coding.
• the device 100 may have a device for setting functional coefficients of a functional specification for adaptation to at least a partial region of the sorted sequence, and the device 120 for encoding may be designed to code the functional coefficients.
• FIG. 1b shows an embodiment of an apparatus 150 for decoding a sequence of samples of an audio signal, each sample having an origin position within the sequence.
• the apparatus 150 comprises means 160 for receiving a sequence of encoded samples, each encoded sample having a sorting position within the sequence of encoded samples, and the means 160 further adapted to receive information about a relationship between the source and sort positions of the samples is.
• the apparatus 150 further comprises means 170 for decoding the samples and the information on the relationship between the source and sort positions, and further comprising means 180 for sorting the samples based on the Information about the relationship between the source and sort positions so that each sample has its original position.
• means 160 may be configured to receive to receive the information about the relationship between the source and sort positions as an index permutation. Further, the receiving means 160 may be configured to receive the information on the relationship between the originating and sorting positions as the inversion chart.
• the means 170 for decoding may be configured in embodiments to decode the encoded samples or the information about the relationship between the source and sort positions with entropy and subsequent difference decoding or only one entropy decoding.
• the means 160 for receiving may optionally receive coded coefficients of a prediction filter and the means for decoding 170 may be configured to decode the coded coefficients, the apparatus 150 further comprising means for predicting samples or relationships between the originating and sorting positions based on may have the coefficient.
• the means 160 may be configured to receive to further receive a residual signal corresponding to a difference between the samples and an output of the prediction filter, and the means 170 for decoding may further be configured to be based on the residual signal to adjust the samples.
• the device 170 can optionally decode the residual signal with an entropy decoding.
• the device 160 for receiving could also receive function coefficients of a function specification, and the device 150 could also have a device for adapting a function specification to at least one subarea of the sorted sequence and the input device.
• Direction 170 for decoding could be designed to decode the function coefficients.
• Fig. 2a shows an embodiment of an apparatus 200 for encoding a sequence of samples of an information signal, each sample having an origin position within the sequence.
• the apparatus 200 includes means 210 for sorting the samples by size to obtain a sorted sequence of samples, each sample having a sorting position within the sorted sequence.
• the apparatus 200 further comprises a means 220 for setting function coefficients of a function specification for adaptation to at least a subarea of the sorted sequence, and means 230 for coding the function coefficients, the samples outside the subarea, and information about a relationship between the originating and sorting positions of the samples.
• the apparatus 200 may further comprise preprocessing means configured to perform filtering, time / frequency transformation, prediction or multi-channel redundancy reduction to generate the sequence of samples.
• the information signal may include an audio signal.
• the means 230 for encoding may be configured to encode the information about the relationship between the source and sort positions as an index permutation. Further, the means 230 for encoding may be configured to encode the information about the relationship between the source and sort positions as an inversion chart.
• means 220 for encoding may also be configured to encode the sorted samples, the information on the relationship between the source and sort positions with difference and subsequent entropy coding, or only entropy coding.
• Means 230 for coding could also be used be formed to determine and encode coefficients of a prediction filter based on the samples, a permutation or an inversion table.
• the means 230 for encoding may be further configured to encode a residual signal corresponding to a difference between the samples and an output of a prediction filter.
• the means 230 for coding may in turn be adapted to code the residual signal with an entropy coding.
• Fig. 2b shows an embodiment of a device 250 for decoding a sequence of samples of an information signal, each sample having an origin position within the sequence.
• the apparatus 250 comprises means 260 for receiving coded function coefficients, sorted samples, and information about a relationship between a sort position and the origin position of samples.
• the apparatus 250 further comprises a means 270 for decoding samples and means 280 for approximating samples based on the function coefficients at least in a partial region of the sequence.
• the apparatus 250 further comprises means 290 for reordering the samples and the approximated portion based on the information about the relationship between the source and sort positions so that each sample has its origin position.
• the information signal may include an audio signal.
• the means 260 for receiving may be arranged to receive the information about the relationship between the originating and sorting positions as an index permutation. Further, the means 260 for receiving may be adapted to receive the information about the relationship between the originating and sorting positions as the inversion chart.
• the device 270 may optionally decode the sorted samples or the information about the relationship between the source and sort positions with entropy and subsequent difference decoding or only entropy decoding.
• the means 260 for receiving may be further adapted to receive coded coefficients of a prediction filter and the means for decoding 270 may be configured to decode the coded coefficients, the apparatus 250 further comprising means for predicting samples based on the coefficients can.
• the means 260 for receiving may be configured to receive a residual signal corresponding to a difference between the samples and an output of the prediction filter or the approximation means 280, and the decoding means 270 may be configured to be based on to adjust the samples to the residual signal.
• the decoding device 270 may optionally decode the residual signal with entropy decoding.
• Fig. 3a shows an apparatus 300 for encoding a sequence of samples of an information signal, each sample having an origin position within the sequence.
• the apparatus 300 includes means 310 for sorting the samples by size to obtain a sorted sequence of samples, each sample having a sorting position within the sorted sequence.
• the apparatus 300 further comprises means 320 for generating a sequence of numbers depending on a relationship between the source and sort positions of the samples and for determining coefficients of a prediction filter based on the sequence of numbers.
• the apparatus 300 further comprises means 330 for encoding the sorted samples and the coefficients.
• the apparatus 300 may further comprise preprocessing means configured to perform filtering, time / frequency transform, prediction or multi-channel redundancy reduction to generate the sequence of samples.
• the information signal may comprise an audio signal.
• the means 320 for generating the number sequence may be designed to generate an index permutation.
• the means 320 for generating the sequence of numbers may generate an inversion chart.
• the means 320 for generating the number sequence may be adapted to further generate a residual signal corresponding to a difference between the sequence of numbers and a prediction order predicted on the basis of the coefficients.
• the means 330 for coding may be adapted to code the sorted samples according to difference and subsequent entropy coding or only one entropy coding.
• the means 330 for coding may further be configured to code the residual signal.
• Fig. 3b shows an embodiment of an apparatus 350 for decoding a sequence of samples of an information signal, each sample having an origin position within the sequence.
• the apparatus 350 comprises means 360 for receiving coefficients of a prediction filter and a sequence of samples, each sample having a sorting position.
• the apparatus further comprises means 370 for predicting a number sequence based on the coefficients, and means 380 for reordering the sequence of samples based on the sequence of numbers so that each sample has its origin position.
• the information signal may comprise an audio signal.
• means 370 for predicting the sequence of numbers may predicate an index permutation as a sequence of numbers.
• the means 370 for predicting the number sequence could also have an inversion table as Predict number sequence.
• the means 360 for receiving may further be adapted to receive a coded residual signal and the means 370 for prediction may be arranged to take into account the residual signal in the prediction of the sequence of numbers.
• the apparatus 350 may further comprise means for decoding configured to decode samples following entropy and then differential decoding or only entropy coding.
• Fig. 4a shows an embodiment of an apparatus 400 for encoding a sequence of samples, each sample having an origin position within the sequence.
• the apparatus 400 includes means 410 for sorting the samples by size to obtain a sorted sequence of samples, each sample having a sorting position within the sorted sequence.
• the apparatus 400 further comprises means 420 for encoding the sorted samples and encoding a sequence of numbers having information about the relationship between the source and sort positions of the samples, each element unique within the sequence of numbers, and wherein the means 420 is for Encoding assigns a number of bits to an element of the sequence of numbers such that the number of bits allocated to a first element is greater than the number of bits allocated to a second element if fewer elements exist prior to encoding the first element already coded before the coding of the second element.
• the means 420 for coding may be designed to code a number sequence of length N and to code a number of X elements simultaneously, wherein the number of X elements G are assigned to bits in accordance with FIG.
• brackets below show that the value in parentheses is rounded up to the next larger integer.
• means 420 for encoding may be configured to encode a number sequence of length N, where X is a number of already encoded elements of the sequence of numbers, where the bits are assigned to the next element of the sequence of numbers G,
• Fig. 4b shows an embodiment of a device 450 for decoding a sequence of samples, each sample having an origin position within the sequence.
• the apparatus 450 comprises means 460 for receiving a coded sequence of numbers and a sequence of samples, each sample having a sorting position.
• the apparatus 450 further comprises means 470 for decoding a decoded number sequence having information about a relationship between the originating and sorting positions based on the encoded sequence of numbers, each element unique within the decoded sequence of numbers and the means 470 for decoding an element of the sequence of numbers allocates a number of bits such that the number of bits allocated to a first element is greater than the number of bits allocated to a second element if fewer elements were already decoded prior to the decoding of the first element than before the coding of the second element.
• the apparatus 450 further includes means 480 for reordering the sequence of samples based on the decoded sequence of numbers such that each sample within the decoded sequence has its origin position.
• means 470 for decoding may be configured to decode a number sequence of length N and to simultaneously decode a number of X elements, wherein the number of X elements are assigned to G bits according to FIG.
• the means 470 for decoding may further be configured to decode a number sequence of length N, where X is a number of already encoded elements of the sequence of numbers, where the next element of the sequence of numbers G bits are assigned according to FIG.
• FIG. 5a shows waveforms of an audio signal 505 (large amplitudes), a permutation 510 (middle amplitudes) and an inversion table 515 (small amplitudes).
• the permutation 510 and the inversion table 515 are shown again in a different scale for the sake of clarity.
• a correlation between the audio signal 505, the permutation 510 and the inversion table 515 can be recognized from the progressions illustrated in FIGS. 5a, b.
• the correlation transfer of the input signal to the permutation or inversion chart can be clearly recognized.
• a permutation coding can take place by creating inversion tables, which are subsequently entropy-coded. It can be seen from FIGS. 5a, b that due to the correlations, a prediction of the permutation or the inversion tables is also possible, wherein the respective residual signal can be entropy-coded in the case of lossless coding, for example.
• the prediction is possible because a correlation present in the input signal is transferred to the resulting permutation or inversion table, cf. Fig. 5a, b.
• FIR Finite Impulse Response
• HR Infinite Impulse Response
• the coefficients of such a filter are then selected so that, for example based on a residual signal at the input of the filter, at its output the original output signal is present or can be output.
• the corresponding coefficients of the filter and the residual signal can then be transmitted more cost-effectively, ie with fewer bits or transmission rate than the original signal itself.
• the original signal In a receiver, or a decoder then the original signal, based on the transmitted coefficients and optionally a residual signal predicted or reconstructed.
• the number of coefficients or the order of the prediction filter determines, on the one hand, the bits necessary for transmission and, on the other hand, the accuracy with which the original signal can be predicted or reconstructed.
• the inversion tables are an equivalent representation of the permutation, but are more suitable for entropy coding. For a lossy coding, it is also possible to perform the resorting only incomplete, thus saving on data traffic.
• FIG. 6 shows an embodiment of an encoder 600.
• pre-processing 605 of the input data may be performed (eg, time / frequency transformation, prediction, stereo-redundancy reduction, band-limiting filtering, etc.).
• the preprocessed data is then sorted 610, resulting in sorted data and a permutation.
• the sorted data can then be processed further.
• be processed or coded 615 for example, here can be a difference coding.
• the data may then be entropy encoded 620 and provided to a bit multiplexer 625.
• the permutation can likewise first be processed or coded 630, for example by determining an inversion table with possibly subsequent prediction, whereupon an entropy coding 635 can also be carried out before the entropy-coded permutation or inversion chart is fed to the bit multiplexer 625.
• the bit multiplexer 625 then multiplexes the entropy-coded data and the permutation into a bitstream.
• FIG. 7 shows an exemplary embodiment of a decoder 700 which, for example, receives a bit stream according to the coder 600.
• the bit stream is then firstly demultiplexed in a bit stream demultiplexer 705, whereupon encoded data is fed to entropy decoding 710.
• the entropy-decoded data may then be included in a decode of the sorted data 715, e.g. in a differential decoding, be further decoded.
• the decoded sorted data is then provided to a resort 720.
• the encoded permutation data are further supplied to an entropy decoding 725, which may be followed by a further decoding of the permutation 730.
• the decoded permutation is then also supplied to the resort 720.
• the resort 720 may then output the output data based on the decoded permutation data and the decoded sorted data.
• Embodiments may further include a coding system having 3 modes of operation.
• a mode 1 could allow high compression rates with the help of a psychoacoustic consideration of the input signal.
• a mode 2 could allow medium compression rates without psychoacoustics, and a mode 3 might allow for lower compression rates, but with lossless encoding, see also TiIo Wik, Dieter Weninger: Lossless audio coding with sorted time values and connection to filter bank based coding methods, October 2006.
• Fig. 8 shows a further embodiment of a coder 800.
• Fig. 8 shows the block diagram of a coder 800 or a coding method for modes 1 and 2.
• MDCT Modified Discrete Cosinus Transform
• the spectral lines are sorted by the magnitude of their amplitudes 810 (sorting). Since the resulting sorted spectrum has a relatively simple waveform, in embodiments it can be easily approximated by a function specification by curve fitting 815 (Curve Fitting), see, eg, Draper, N.R and H. Smith, Applied Regression Analysis, 3rd Ed. John Wiley & Sons, New York, 1998.
• curve fitting 815 Curve Fitting
• a reordering rule 820 can now be found and into the bit stream be written, which includes the smallest possible amount of data.
• mode 1 can be accomplished by, for example, a run-length encoding 820 and for mode 2 by a special permutation encoder 820 that can operate using an inversion table.
• the data of the run-length coding or of the permutation coder 820 are then additionally coded by an entropy coding method or entropy coder 830 and finally written into the bitstream including some additional information, eg the coefficients of the above-mentioned functional rule, indicated by the bitstream formatter 835.
• FIG. 8 also shows a block 825 which monitors the bit rate produced in the encoding process and, if necessary, provides feedback to the psychoacoustic model if the data rate is still too high.
• FIG. 8 shows that a psychoacoustic model 840 for bitrate control can be activated, for example, only for mode 1, and in mode 2 this control option in favor of the coding quality can be dispensed with.
• operation mode 1 a higher compression rate is achieved than in the other two operating modes.
• lines of the frequency spectrum are purposefully set to zero by means of a psychoacoustic observation 840 of the input signal, or alternatively elements of the index permutation are excluded from the resorting in order to be able to save data in the transmission of the resorting rule 820.
• operating mode 2 the frequency spectrum is completely restored, there are only small errors due to small inaccuracies of the curve approximation 815.
• the operating mode 2 can be extended with the addition of a residual signal to a lossless mode. Both in mode 1 and mode 2, the entire frequency spectrum can be transmitted, ie the data reduction in Mode 1 can only be achieved by a reduced re-sort rule 820.
• FIG. 9 shows a further embodiment of a decoder 900 or a decoding process of modes 1 and 2, which essentially passes back through the steps of the encoding or of the encoder 800.
• the bitstream is first unpacked by a bit stream demultiplexer 905 and decoded in an entropy decoder 910. From the decoded function coefficients of a function specification, the function or spectral function can then be restored by an "inverse curve fitting" block, ie an inverse curve fit 915, and fed to a resorter 920.
• the resorter 920 also receives a permutation from a permutation decoder 925, which decodes the permutation on the basis of the entropy-coded permutation, and the resorter 920 can then restore its spectral lines to their original order by means of the permutation and the spectral function reconstructed with the help of the transmitted function coefficients Finally, the reconstructed spectrum is transformed by an inverse transformation 930 eg inverse MDCT, transformed back into the time domain.
• an inverse transformation 930 eg inverse MDCT
• FIG. 10 shows an example of a frequency spectrum of an audio signal with 1024 frequency lines and its approximated spectrum, where the original and the approximation are almost identical.
• 11 shows the associated sorted spectrum and its approximation. It is clear that the sorted spectrum can be more easily and accurately approximated by a functional specification than the original spectrum.
• the spectrum of FIG. play into, for example, 5 areas (partitions or English partitions), which are shown in FIG. 11, the area 3 being characterized by a straight line and the areas 2 and 4 having correspondingly suitable functions (eg polynomials, exponential functions, etc.). ).
• the number of amplitude values in the ranges 1 and 5 can be chosen very small in embodiments, eg 3, but since these are enormously important for the sound quality, they should either be very accurately approximated or transmitted directly.
• FIG. 10 additionally shows the approximated and sorted back spectrum, wherein it is easy to see that the reconstructed spectrum comes very close to the original spectrum.
• the resorting results in a number sequence of the spectral line indices, which represents a permutation of the index quantity.
• the number sequence of these resorted indexes can be transmitted directly, whereby relatively large amounts of data can arise, which, since they are completely uniform, can not be reduced by an entropy coding.
• this sequence of numbers is logically unsorted to map to an unequally distributed sequence, in embodiments an inversion table formation can be applied to the indices, which represents a bijective, ie unambiguously reversible image, and delivers a not uniformly distributed result, see. eg Donald E.
• A is sorted by the magnitude of V 1 so that V 1 form a monotone decreasing sequence.
• the X 1 become thereby an unsorted number sequence, thus a permutation of the original X 1 .
• a 1 ⁇ (5,8), (9,6), (1,5), (8,4,5), (2,3), (6,2,3), (4,2), (7,2 ), (3,1) ⁇
• the inversion of the inversion table returns the original sequence of numbers:
• a differential encoding would be used, e.g. in Ziya Arnavut: Permutations Techniques in Lossless Compression, Dissertation, 1995, or other post-processing (e.g., prediction) that reduce entropy is conceivable.
• Embodiments of the present invention operate on a completely different principle than already existing systems. By avoiding the computational steps of quantization, resampling and low-pass filtering as well as the optional omission of a psychoacoustic view, embodiments can save on computational complexity.
• the quality of the coding for the mode 2 depends exclusively on the quality of the approximation of the functional specification to the sorted frequency spectrum, while for the mode 1 the quality is mainly determined by the psychoacoustic model used.
• bitrate of all modes depends for the most part on the complexity of the resort rule to be transferred.
• the bit rate scalability is given in a wide range, from high compression to lossy Free coding at higher data rates, every gradation is possible. Due to the principle of operation, the full frequency bandwidth of the signal can be transmitted even at relatively low bit rates.
• the low requirements for the computing power and the memory space allow an application and implementation of embodiments not only on conventional PCs, but also on portable devices.
• Embodiments may additionally provide for the transmission of an error or residual signal, thereby increasing the quality of modes 1 and 2 and extending mode 2 even to a lossless mode. Furthermore, a transmitted error signal could allow intelligent retirement for the frequency lines excluded from resorting in the mode 1, thus further improving the quality of this mode.
• SBR Spectral Band Replication
• a psychoacoustic consideration tailored specifically to the errors arising in the approximation could increase the quality and reduce the bit rate in further embodiments. Since the principle of resorting and subsequent curve approximation is not dependent on signals from the frequency domain, other embodiments may also be used in the time domain for the mode 2. Since the modes 2 and 3 dispense with the use of a psychoacoustic analysis, embodiments can also be used outside the audio coding.
• Embodiments may further provide optimized processing of stereo signals adapted to the particularities of this method and may thus further reduce the bit consumption and the computational outlay over a dual mono-encoding.
• Exemplary embodiments make use of a sorting model.
• a sorting of the data to be coded takes place. This, on the one hand, causes an artificial correlation of the data, which makes the data easier encode.
• the sorting results in a permutation of the original positions of the time values.
• a reordering rule permutation
• the original problem of merely encoding the time values is now broken down into two subproblems, that is to say a coding of the sorted time values and a coding of the sorting rule.
• 11 illustrates the scheme of a so-called sorted-lossless coding, for example, an audio signal is mapped to a signal with a stronger correlation, and then the sorted time values and a sorting rule are coded.
• SOLO sorted lossless
• SOLO sorted lossless
• difference coding as the name implies, not the actual value but the difference of successive values is coded. If the differences are smaller than the original values, a higher compression can be achieved.
• Curve fitting is a technique that attempts in exemplary embodiments to adapt a given mathematical model function as best as possible to data points, in this case the sorted time values.
• the effectiveness of curve fitting is essentially determined by the shape of the curves to be described. What is certain is that, depending on the type of sorting, it is a monotonously decreasing or monotonically increasing waveform.
• Figures 12 and 13 show two representative waveforms of sorted time values. Noteworthy is the non-uniform waveform in Fig. 13. Such curves, which occur in about 40% (related to a selection of different audio signals) of the cases, can usually not be described very well by a Curve Fitting.
• CpC 29 Zl 15 Zl 2 are elements of the set of real numbers and can be determined eg with the Nelder-Mead Simplex algorithm, cf. NELDER, YES; MEAD, RA: A Simplex Method for Function Minimization. Computer Journal, Vol. 7, p. 308-313, 1965.
• This algorithm is a method for optimizing nonlinear functions of multiple parameters. Similar to the Regula-falsi method with step size control, the tendency of the values towards the optimum is approximated.
• the Nelder-Mead Simplex algorithm converges roughly linearly and is relatively simple and robust.
• the function f cfl has the advantage that it can adapt very flexibly to a whole series of curves. The disadvantage, however, is that relatively much page information (four coefficients) is needed. It is also noticeable that parts of the sorted curves, for example the middle section of FIG. 12, could well be described by a polynomial of first degree (straight line) and only two real coefficients a, b would be needed. Therefore, a second function should alternatively be used:
• a curve fit over the total number of sorted time values of a block is certainly too inaccurate. Therefore, it would be useful to divide the block into several smaller partitions. If, however, the block is decomposed into too many parti- tions, which are described by the functions f c ⁇ and f Cf ⁇ , one needs many functional coefficients . Therefore, in one embodiment, with a fixed total block length of 1024 time values, a division is made into 4 partitions of 256 time values each. In order to be able to decide per partition whether f cfl or f ci2 is better suited for curve fitting, an adequate decision criterion is required. The decision criterion should, on the one hand, be easy to identify and, on the other hand, be resilient.
• Fig. 14 the operation of Curve fitting is shown.
• the first and fourth partitions are described by fcf2 and the second and third partitions are written by fcfl.
• GZIP uses for compression the Deflate algorithm, which is a combination of LZ77 (see Ziv, Jacob; Lempel, Abraham: A Universal Algorithm for Sequential Data Compression, IEEE Transactions on Information Theory, Vol. IT-23, No. 3, May 1977 ) and Huffman coding (see Huffman, David A.: A Method for the Construction of Minimum Redundancy Codes, Proceedings of the IRE, September, 1952).
• the ZIP file format uses a similar algorithm for compression.
• Another universal method is BZIP2. Prior to the actual encoding of the data, this is preceded by the Burrows-Wheeler Transform (BWT) (see Burrows, M. Wheeler, D .: A block sorting lossless data compression algorithm, Technical Report 124, Digital Equipment Corporation, 1994). instead of.
• BWT Burrows-Wheeler Transform
• BZIP2 also uses a Huffman encoding. These programs can be applied to any data such as text, program code, audio signals, etc. apply. However, due to their mode of operation, these methods achieve much better compression in text than in audio.
• a direct comparison of GZIP and the SHORTEN compression method specializing in audio signals confirms this, see following table. The default settings were used for the test.
• FIG. 16 exemplifies processing steps of most lossless audio compression algorithms.
• the illustration in FIG. 16 shows a block diagram in which the audio signal is first subjected to block formation or an English-language block diagram. Subsequently, an intra-channel decorrelation block or "intra-channel decorrelation” block individually decorrels the signal, for example by differential coding. In an entropy coding block or engl. "Entropy Coding" block finally entropy-codes the signal, see also Hans, Mat; Schafer, Ronald W .: Lossless Compression of Digital Audio, IEEE Signal Processing Magazine, June 2001.
• the data to be processed are divided into signal sections (frames) x (n) ⁇ Z (Z corresponds to the set of integers) of a certain size.
• This is followed by a decorrelation step, in which an attempt is made to remove the redundancy from the signal as well as possible.
• the signal e (n) CZ resulting from the decorrelation step is entropy-coded. So far, there are two principal approaches to the decorrelation step. Most lossless audio coding methods use a kind of linear prediction to remove the redundancy from the signal (predictive modeling).
• lossless audio coding methods are based on a lossy audio coding method in which, in addition to the lossy data, the residual or error signal (even residual signal) is additionally coded to the original signal (Lossy Coding Model). Subsequently, the different approaches are to be considered more closely.
• Linear Predictive Coding is widely used in digital speech signal processing. Their importance lies not only in the high efficiency, but also in the relatively low computational complexity.
• the basic idea of the prediction is a value x (n) of previous values x (nl), x (n-2),. , , to predict x (np). If p preceding values are used for prediction, this is called a pth order predictor.
• the prediction coding methods used in lossless audio coding usually have the basic structure shown in FIG. A ⁇ z) and B ⁇ z) stand for z-transformation polynomials (see Mitra, Sanjit K .: Digital Signal Processing, New York: McGraw-Hill, 2001, pp.
• FIG. 17 shows an embodiment of a structure of prediction coding.
• FIG. 17 shows an HR filter structure with a feedforward branch with filter coefficients A (z), a negative feedback branch with filter coefficients B (z), and a quantization Q.
• IIR predictors are much more complex, but in some cases they can achieve better encoding gains than FIR pseudopods (see Craven, P., Law, M., Stuart J: Lossless Compression Using HR Prediction Filters, Kunststoff: 102nd AES Conv , 1997).
• FIR pseudopods see Craven, P., Law, M., Stuart J: Lossless Compression Using HR Prediction Filters, Kunststoff: 102nd AES Conv , 1997.
• Fig. 18 shows an embodiment of a structure of reconstruction in prediction coding.
• Fig. 18 shows an embodiment as IIR filter structure with a feedforward branch with filter coefficients 5 (z), a negative feedback branch with filter coefficients A ⁇ z), and a quantization Q. Behind Fig. 18, the equation hides
• the predictor coefficients are redetermined and transmitted each time for each signal portion to be processed.
• the adaptive determination of the coefficients ak of a pth order predictor can be carried out either with the covariance method or with the autocorrelation method which uses the autocorrelation function.
• the autocorrelation method obtains the coefficients by solving a linear system of equations of the following form:
• the equation can be very effectively solved with the Levinson-Durbin algorithm because of the matrix properties of R (see Yu, R., Lin, X., Ko, CC: A Multi-Stage Levinson-Durbin Algorithm., IEEE Proc, Vol 218-221, November 2002, pp. 218-221.
• a division of the time values into blocks of size N is made. Assuming one would like to use a second-order predictor to predict the time values from the current block n, then the problem arises as to how to deal with the first two values from block n. One can either use the last two values from the previous block n-1 to predicate them, or not predict the first two values of block n and leave them in their original form. If values from a previous block n-1 are used, block n is only decodable if block n-1 was successfully decoded. However, this would lead to block dependencies and contradict the principle of treating each block (frame) as a stand-alone decodable unit.
• the first p values are left in their original form, they will be called engl.
• Warmup or warm-up values (see Fig. 19) of the predictor referred. Because the warmup usually has different proportions and statistical properties than the residual signal, it is usually not entropy-coded.
• Fig. 19 shows an embodiment of warmup values of a prediction filter.
• unchanged input signals are shown; in the lower area, warm-up values and a residual signal are shown.
• Another possibility for realizing a prediction is not to re-determine the coefficients for each signal section, but always to use fixed predictor coefficients. If you always use the same coefficients, you speak of a fixed predictor.
• AudioPaK see Hans, Mat; Schafer, Ronald W .: Lossless Compression of Digital Audio, IEEE Signal Processing Magazine, June 2001, pp. 28-31
• AudioPaK splits the audio signal into independent decodable sections. Usually, multiples of 192 samples (192, 576, 1152, 2304, 4608) are used. For decorrelation, a FIR predictor with fixed integer coefficients is used (fixed predictor). This FIR predictor was first used in SHORTEN (see Robinson, Tony: SHORTEN: Simple lossless and nearloss waveform compression, Technical Report CUED / FINFENG / TR.156, Cambridge University Engineering Department, December 1994, pp. 3-4). Internally, the Fixed Predictor has four different predictive models.
• the equation is polynomial approximation or prediction methods.
• degree p-1 the preceding p samples x (nl), x (n-2), ..., x (np) 'can be described. If one now evaluates this polynomial at the position n, one obtains the predicted value x (n). Graphically, this can be illustrated as shown in FIG. Fig. 20 shows an embodiment of a predictive model, in a polynomial predictor.
• AudioPaK uses a Rice coding. Since the values of the residual signal ei (n) GZ, but the Rice coding works with values from No, an initial mapping of the residual values ei (n) to No is made.
• the Rice parameter k is determined per block (frame) and assumes values 0,1, ..., (b-1). Where b represents the number of bits per audio sample, k is determined by the following equation
• Framesize is the number of samples per frame and AbsError is the sum of the absolute values of the residual signal.
• Further representatives of predictive modeling are SHORTEN (see Robinson, Tony: SHORTEN: Simple lossless and nearlossless waveform compression, Technical Report CUED / FINFENG / TR.156, Cambridge University Engineering Department, December 1994), FLAC (see Coalson, Josh: FLAC).
• MPEG-4 Audio Lossless Coding MPEG-4 ALS
• Monkey 's Audio see Ashland, Matthew T .: Monkey' s Audio - http://www.monkeysaudio.com) /index.html).
• the second way to implement a lossless audio coding method is to set up a lossy audio coding method.
• a representative of the Lossy Coding Model is LTAC, which uses the abbreviation LTC (Lossless Transform Coding) instead of LTAC (Lossless Transform Audio Compression), see Liebchen, Tilman; Purat, Marcus; NoIl, Peter: Lossless Transform Coding of Audio Signals. Kunststoff, Germany: 102nd AES Convention, 1997.
• LTC Low Transform Coding
• LTAC Lossless Transform Audio Compression
• Fig. 21 shows a block diagram of a structure of a LTAC (Lossless Transform Coding) encoder.
• the encoder comprises a "DCT" block to transform an input signal x (n) into the frequency domain, followed by quantization Q.
• the quantized signal c (n) can then be transformed back into the time domain via an "IDCT” block where it can then be subtracted, quantized by another quantizer Q, from the original input signal.
• the residual signal e (n) can then be transmitted entropy-coded.
• the quantized signal c (n) can also be coded via an entropy coding, which can select from various codebooks according to FIG. 21.
• the time values x (n) are transformed into the frequency domain by an orthogonal transformation (DCT - discrete cosine transformation).
• DCT orthogonal transformation
• the spectral values are then quantized c (k) and entropy-coded.
• the quantized spectral values c (k) are additionally transformed back with the inverse discrete cosine transformation (IDCT) and quantized again y (n).
• e (n) is entropy-coded and transmitted.
• the decoder one can again obtain y (n) from c (k) through the IDCT with subsequent quantization.
• MPEG-4 Scalable Lossless Audio Coding (MPEG-4 SLS) (see Geiger, Ralf, Yu, Rongshan, Herre, Jurgen, Rahardja, Susanto, Kim, Sang-Wook ; Lin, Xiao; Schmidt, Markus: ISO / IEC MPEG-4 High-Definition Scalable Advanced Audio Coding, Paris: 120th AES Convention, May 2006). It combines lossless audio encoding, lossy audio coding and scalable audio coding functionality.
• MPEG-4 SLS is backward compatible to MPEG-4 Advanced Audio Coding (MPEG-4 AAC) at bitstream level (see ISO / IEC JTC1 / SC29 / WG11: Coding of Audiovisual Objects, Part 3. Audio, Subpart 4 Time / Frequency Coding Standard 14496-3, 1999).
• Fig. 22 shows a block diagram of an MPEG-4 SLS (Scalable Lossless Audio Coding) SLS encoder.
• the audio data are first of all analyzed using an IntMDCT (Integer Modified Discrete Cosine Transform) (see Geiger, Ralf, Sporer, Thomas, Koller, Jürgen, Brandenburg, Karlheinz: Audio Coding Based on Integer Transforms, New York: lllnd AES Conv., 2001) transformed the frequency domain and then further processed by a Temporal Noise Shaping (TNS) and a middle / side channel coding (Integer AAC Tools / adaptation). Everything encoded by the AAC encoder is then removed from the IntMDCT spectral values by an error mapping / error mapping. What remains is a residual signal which is subjected to entropy coding. For entropy coding, a BPGC (Bit-Plane Golomb Co-de), CBAC (Context-Based Arithmetic Code) and Low Energy Mode / Low Energy Mode are used.
• IntMDCT Integer Modified Discrete Cosine Transform
• stereophony The sound transmission over two or more channels is called stereophony.
• stereo is usually used exclusively for two-channel pieces. If there are more than two channels, this is called multi-channel sound.
• This master thesis is limited to signals with two channels, for which the term stereo signals is used synonymously.
• One way to handle stereo signals is to encode both channels independently. In this case, one speaks of independent stereo coding (Independent Stereo Coding).
• Independent Stereo Coding Apart from "pseudo-stereo" versions of old mono recordings (both channels identical) or two-channel sound in television (independent channels), stereo signals usually have both differences and similarities (redundancy) between the two channels. If you can identify the similarities and transfer them only once for both channels, you can reduce the bit rate.
• NINT here means rounding to the nearest integer with respect to zero.
• lossless audio coding methods also use LS coding or RS coding (see Coalson, Josh: FLAC - Free Lossless Audio Codec; http://flac.sourceforge.net).
• LS coding or RS coding see Coalson, Josh: FLAC - Free Lossless Audio Codec; http://flac.sourceforge.net.
• FIG. 23 shows a stereo-redundancy reduction (SRR) after decorrelation of individual channels
• FIG. 24 shows a stereo-redundancy reduction before decorating individual channels. Both methods have specific advantages as well as disadvantages. In the following, however, only method 2 should be used.
• the determined coefficients a z are usually floating-point values (real numbers), which can only be represented with finite precision in digital systems. So there must be a quantization of the coefficients a z . However, this can lead to larger prediction errors and has to be considered in the generation of the residual signal. Therefore, it makes sense to control the quantization via an accuracy parameter g. If g is large, finer quantization of the coefficients takes place and more bits are needed for the coefficients. If g is small, a coarser quantization of the coefficients takes place and fewer bits are needed for the coefficients. In order to be able to realize a quantization, the largest coefficient a max is initially determined
• the residual signal e (n) to be transmitted is determined
• FIG. 25 shows the relationship between predictor order and total bit consumption. It will be appreciated that as the order increases, the residual signal will require fewer and fewer bits for encoding. However, the data rate for the page information (quantized prediction coefficients and warmup) increases continuously, causing the total data rate to rise again at a certain point. As a rule, a minimum is reached at Kp ⁇ 16. In Fig.
• Fig. 26 shows an illustration of the relationship between quantization parameter g and total bit consumption. If one considers the total data rate as a function of the quantization parameter g (see FIG. 26), the bit consumption for the residual signal continuously drops to a certain value. From here, a further increase in quantization accuracy is of no avail. That the necessary number of bits for the residual signal remains approximately constant. The overall data rate initially decreases continuously, but then increases again due to increasing page information for the quantized predictor coefficients. In most cases an optimum results at 5 ⁇ g ⁇ 15. In Fig.
• g ll.
• the newly gained insights should now be used to give an algorithm for lossless linear prediction in a simplified MATLAB code representation (see lpc ()).
• MATLAB is a commercial mathematical software designed for calculations with matrices. Hence the name MATrix LABoratory. Programming is done under MATLAB in a proprietary, platform-independent programming language, which is interpreted on the respective computer.
• some variables are initialized after the limits found in FIGS. 25 and 26. Then the predictor coefficients are determined via the autocorrelation and the Levinson-Durbin algorithm.
• the core of the algorithm is formed by two nested for loops.
• the outer loop passes over the predictor order p.
• the inner loop passes over the quantization parameter g.
• the quantization of the coefficients, the calculation of the residual signal and an entropy coding of the residual signal take place. Instead of a complete entropy coding of the residual signal, an estimate of the bit consumption which might possibly be carried out faster would also be conceivable. Finally, you secure the variant with the lowest bit consumption.
• An example of a MATLAB code follows:
• the transfer function is a mathematical description of the behavior of a linear time-invariant system that has an input and an output.
• the frequency response describes the behavior of a linear time-variant system, comparing the output with the input and recording it as a function of frequency.
• X 3 (/ ⁇ ) 3X ( ⁇ -1) -3X ( ⁇ -2) + X ( ⁇ -3)
• Fig. 27 shows an illustration of a magnitude frequency response of a fixed predictor as a function of its order p.
• the effect of the different predictive orders becomes clear on the basis of a consideration of their frequency response (see Fig. 27).
• the residual signal corresponds to the input signal.
• Increasing the order leads on the one hand to a greater attenuation of the low-frequency signal components, but on the other hand to an increase in the high-frequency signal components.
• the frequency axis was normalized for the representation at half the sampling frequency, resulting in 1 at half the sampling frequency (here 22.05 kHz).
• the following table shows how often which order was selected as the best, summed over the entire audio file. For the creation of this table, a constant block length of 1024 time values was taken.
• the newly gained insights should be used to specify an algorithm (see fixed ()).
• the maximum and minimum order are set.
• the residual signal with the corresponding bit consumption and the costs for the warmup as a function of the order are determined.
• the best option is chosen.
• the difference coding is invertible.
• Difference coding works optimally if the values to be coded are very close to each other, ie strongly correlated. Sorting the time values places the time values in a strong correlation. Figure 12 already showed the effect of difference coding applied to sorted time values. The matching value of sorted and decorrelated time signal at index 1 (warmup) can be clearly seen. Furthermore, the much smaller dynamic range of the residual signal of the difference coding compared to the sorted time values is striking. Detailed information on FIG. 12 is given in the following table. The difference coding thus represents a simple and efficient method for coding sorted time values.
• Example By way of example, an example of the formation of an inversion table RS and the corresponding generation of the permutation is shown here.
• Lehmer Inversion panel LB Lehmer Inversion panel LB.
• cr ⁇ S ⁇ be a permutation.
• the Lehmer inversion table LB / / w (cr) ⁇ b ⁇ , b 2 , ..., b n ) is then defined as
• a lg orithm 1 ⁇ 1 , ( ⁇ ) 1. Set i ⁇ - n, 1 ⁇ - (n, nl, ..., l) 2. ⁇ (i) ⁇ - l (b, + l)
• the property shown of the elements of the inversion panel LB also applies to the inversion panels RB, RBL and RSL.
• the elements have the following properties
• a / P (Shuffling): Let X 1 , X 2 , ..., X be a number of t numbers to be scrambled.
• Fig. 28 shows an illustration of the relationship between permutation length
• FIGs. 29a-h show a representation of inversion tables in the 10th block (frame) of a noise-like piece.
• Figures 30a-h show a representation of inversion tables in the 20th block (frame) of a tonal piece. It is based on a block size of 1024 time values.
• FIGS. 31 a, b show a representation of a sorting of time values of a permutation of a noise-like piece in the 10th block and of a tonal piece.
• Fig. 31 The right permutation of Fig. 31 is pronounced of a main signal mirrored audio signal. It seems that there is a direct correlation between the audio signal, the sorting rule and even the inversion charts.
• Figures 32a, b and 33a, b show the audio signal of a block, the corresponding permutation at which the x and y coordinates were swapped and the corresponding inversion table LS.
• Fig. 32a shows a part of an audio signal, the corresponding permutation and inversion table LS
• Fig. 32b shows the permutation and the inversion table LS of Fig. 32a enlarged.
• Fig. 33a shows a part of an audio signal, the corresponding permutation and inversion table LS
• Fig. 33b shows the permutation and the inversion table LS of Fig. 33a enlarged.
• Figures 32a, b and 33a, b clearly show the connectivity of the original audio signal, permutation and inversion chart. That is, as the amplitude of the original signal increases, so does the amplitude of the permutation and inversion chart, and vice versa. Very worth mentioning are the amplitude ratios.
• a 16-bit audio signal has one
• bitsLB, bitsLS, bitsRB, bitsRSl bitsLBL, bitsLSL, bitsRBL, bitsRSL] getBitConsumptio ⁇ (residualignalLB, residualsignalLS, residualsignalRB, residualsignalIRS, residualsignalLBL, residualsignalSLS, residualsignalIRBL, residualsignalIRSL); % determine the least expensive variant
• a dynamic bit allocation realized via the inversion tables or Lehmer inversion tables can save approximately IBit per element compared to a conventional binary coding of the permutation. This coding approach thus represents a simple and profitable procedure for the worst case case.
• Fig. 34a shows a probability distribution
• Fig. 34b shows a length of the codewords of a predicted (Fixed Predictor) residual signal of an inversion chart LB.
• Fig. 34a shows the probability distribution of the residual signal of a non-Lehmer Inversion Tafel LB, created by the application of a fixed predictor.
• Golomb or Rice coding is optimally suitable as entropy coding methods (see GOLOMB, SW: Run-length encodings, IEEE Transactions on Information Theory, IT-12 (3), pp. 399-401, June 1966, GALLAGER, Robert G .; VAN VOORHIS, David C: Optimal Source Codes for Geometrically Distributed Integer Alphabets. IEEE Transactions on Infor- Theory, March 1975 and Salomon, David: Data Compression. London: Springer-Verlag, Fourth Edition, 2007, Salomon, David: Data Compression. London: Springer-Verlag, Fourth Edition, 2007.
• Fig. 35a shows a probability distribution
• the residual signals have the property that the value ranges from block to block sometimes vary considerably and many values of the value range do not occur at all. In Fig. 34, for example, this is the case between -25, ..., - 20. This can also be seen in FIG. 35 for values> 350.
• the easiest way to determine the Rice parameter is to test all the Rice parameters in question and choose the parameter with the lowest bit consumption. This is not very complex because the range of values of the Rice parameters to be tested is determined by the bit resolution of the time period. Signal is limited. At a resolution of lBit, a maximum of 16 Rice parameters can be verified. The corresponding bit requirement per parameter can ultimately be determined by means of a few bit operations or arithmetic operations. This procedure of finding the optimal Rice parameter is a bit more complicated than the direct calculation of the parameter, but always guarantees the optimal Rice parameter. In the lossless audio coding method presented here, in most cases this method is used to determine the Rice parameter. In a direct determination of the Rice parameter one can look at the KIELY, A.: Selecting the Golomb parameters in Rice Coding. IPN Progress Report, Vol. 42-159, November 2004 derived parameter limits
• FIG. 36 shows a percentage of a subblock decomposition with a least data load of a forward adaptive Rice coding over a residual signal of a fixed predictor of a piece including a page information for parameters, the total block length is 1024 time values large.
• a partial block decomposition is usually not very profitable. If the Rice parameters are coded, then a decomposition into 32 subblocks is often better than no subblock decomposition (see also the following table).
• sub-block decomposition is usually not advantageous for both uncoded Golomb parameters and encoded Golomb parameters (see Figure 37 and the following table).
• Fig. 37 shows a percentage of a partial block decomposition with the least amount of data of a forward adaptive Golomb coding over the residual signal of a fixed predictor of a piece including side information for parameters, the total block length being 1024 time values.
• the Rice parameters are in principle quantized Golomb parameters, this should not be considered here.
• a backward-adaptive Rice / Golomb coding calculates the parameter from previously encoded characters. For this purpose, the newly encoded characters are entered cyclically in a history buffer. There are two variables to the history buffer. One holds the current level of the History Buffer and the other variable stores the next write position.
• FIG. 38 shows the basic operation of the size 8 history buffer. Initially, the history buffer is initialized to zero, the level is zero and the write index is one (see a)). Then one character at a time is entered in the history buffer and the writing index (arrows) and the filling level are updated (see b) -e)). Once the history buffer has been completely filled, the level remains constant (here 8) and only the writing index is adjusted (see e) -f)).
• the adaptation used for decoding must be synchronized for encoding, otherwise a perfect reconstruction of the data is not possible.
• the reverse buffer does not yet have a good prediction of the parameter. Therefore, a variant is used which calculates a forward-adaptive parameter for the first W values, and only when the history buffer is completely filled are it calculated from this adaptive parameter.
• FIGS. 39 a, b show an illustration of the operation of an adaptation compared with an optimum parameter for the overall block.
• the brighter lines represent the limit range from which the adaptive parameters are used. Simplified, this procedure just described can be represented as in BwAdaptivCodingO.
• e (i) ⁇ Z an image after No.
• a forward-adaptive parameter is determined with which the first W values are coded. If the history buffer is completely filled, the adaptive parameters are used for the further coding.
• a forward adaptive arithmetic coding with the help of a backward adaptive Rice coding is to be developed in addition.
• a histogram of the data to be encoded is created. With this histogram it is possible to generate a code close to the entropy limit by the arithmetic coding.
• the contained characters and their occurrence probabilities must be transmitted. Since the characters in the histogram are arranged in a strictly monotonically increasing manner, a differential coding ⁇ is still available here before a backward-adaptive Rice coding. The probabilities are only encoded backwards-adaptive Rice.
• Fig. 40 shows an embodiment of a forward adaptive arithmetic coding with the aid of a backward adaptive Rice coding.
• the following table shows a comparison of different entropy coding methods applied to the residual signal. Comparison of different entropy coding methods applied to the residual signal of the fixed predictor.
• the following table shows a comparison of different entropy coding methods applied to the residual signal of the difference coding of the sorted time values
• the Golomb parameter requires a slightly higher page information data rate than the Rice parameter. Nevertheless, the backward-adaptive Golomb coding on average represents the best entropy coding method for the residual signals present in SOLO. In very rare cases, it may happen that the adaptation strategy fails and does not give good results. For this reason SOLO uses a combination of a backward adaptive Golomb coding and a forward adaptive Rice coding.
• both the encoder and the decoder require large data structures to hold the data to be processed in memory.
• the first decoded data are available later at a larger block length than at a smaller block length
• the block length is determined by what requirements are imposed on the coding method. If the compression factor is in the foreground, a very large block length may be acceptable. But if a coding method with a low delay time or a small memory consumption is required, then a very large block length is certainly not useful.
• Existing audio coding techniques typically use block lengths of 128 to 4608 samples. At a sampling rate of 44.1 khz, this corresponds to approximately 3 to 104 milliseconds.
• An investigation should provide information how the different decorrelation methods used by SOLO behave at different block lengths. For this purpose, various pieces are encoded at block lengths of 256, 512, 1024 and 2048 samples, and the compression factor F is determined using the respective side information. Of the seven compression factors of one block length, the arithmetic mean is then formed. In Fig. 41 the result of this investigation is shown.
• FIG. 41 shows a representation of the influence of the block size on the compression factor F. It can be clearly seen that the predictors achieve a better compression factor as the block length increases, and this is not so pronounced in the case of the fixed predictor as in the LPC coding method.
• the sorting model decorrelation method has an optimum at a block length of 1024 samples. Since a high compression factor is desirable with as small a block length as possible, a block length of 1024 samples is primarily used in the following. However, SOLO can optionally be operated with a block length of 256, 512 or 2048 samples.
• Fig. 42 shows a diagram for lossless MS encoding.
• the MS decoding inverts the calculation rule of the MS encoding and generates from M and S again the right channel R and the left channel L
• FIG. Fig. 43 shows another illustration of lossless MS encoding.
• LR coding no stereo redundancy reduction
• LS coding left channel and side channel
• MS coding center channel and side channel
• FIG. 44 shows an illustration for selecting a best variant for stereo-redundancy reduction.
• the inventive concept can also be implemented in software.
• the implementation can take place on a digital storage medium, in particular a floppy disk or a CD with electronically readable control signals, which can cooperate with a programmable computer system and / or microcontroller such that the corresponding method is carried out.
• the invention thus also consists in a computer program product with a program code stored on a machine-readable carrier for carrying out the method according to the invention, when the computer program product runs on a computer and / or microcontroller.
• the invention can thus be realized as a computer program with a program code for carrying out the method, when the computer program runs on a computer and / or microcontroller.

Landscapes

• Engineering & Computer Science (AREA)
• Physics & Mathematics (AREA)
• Signal Processing (AREA)
• Health & Medical Sciences (AREA)
• Audiology, Speech & Language Pathology (AREA)
• Human Computer Interaction (AREA)
• Computational Linguistics (AREA)
• Acoustics & Sound (AREA)
• Multimedia (AREA)
• Theoretical Computer Science (AREA)
• Compression, Expansion, Code Conversion, And Decoders (AREA)
• Compression Or Coding Systems Of Tv Signals (AREA)
• Transmission Systems Not Characterized By The Medium Used For Transmission (AREA)
• Reduction Or Emphasis Of Bandwidth Of Signals (AREA)

Applications Claiming Priority (3)

Publications (2)

Family

ID=39283871

Family Applications (1)

Country Status (9)

Families Citing this family (37)

Family Cites Families (22)

• 2007
  • 2007-04-12 DE DE102007017254A patent/DE102007017254B4/de active Active
  • 2007-11-16 CN CN2007800498867A patent/CN101601087B/zh active Active
  • 2007-11-16 AT AT07819842T patent/ATE527655T1/de active
  • 2007-11-16 US US12/514,629 patent/US20100027625A1/en not_active Abandoned
  • 2007-11-16 WO PCT/EP2007/009941 patent/WO2008058754A2/fr active Application Filing
  • 2007-11-16 KR KR1020097011487A patent/KR101122573B1/ko active IP Right Grant
  • 2007-11-16 JP JP2009536665A patent/JP5200028B2/ja active Active
  • 2007-11-16 EP EP20070819842 patent/EP2054884B1/fr active Active
• 2009
  • 2009-06-05 HK HK09105056.1A patent/HK1126568A1/xx unknown

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events